• Designed and deployed advanced AI services integrating LLMs, retrieval-augmented generation (RAG), embeddings, and vector databases, delivering real-time financial insights, predictive recommendations, and automated client advisory workflows.

• Architected scalable machine learning data pipelines with Apache Spark, AWS Glue, and Kafka, enabling continuous ingestion, feature engineering, and low-latency model inference on large-scale streaming datasets.

• Developed and containerized microservices for model serving and inference using Docker, Kubernetes, and CI/CD, achieving 99.9% system uptime, rapid iteration cycles, and cost-efficient deployment at scale.

• Implemented observability, monitoring, and experiment tracking solutions with Prometheus, Grafana, and MLflow to measure key performance indicators (latency, drift, cost), improve model reliability, and support audit readiness.

• Built end-to-end healthcare AI platforms using TensorFlow and PyTorch, predictive diagnostic models, recommendation systems, and clinical decision-support tools that improved patient outcome predictions.

• Developed retrieval-augmented generation (RAG) and semantic search pipelines with custom embeddings, automating medical record retrieval and powering context-aware, natural language responses for patient interactions.

• Designed data workflows for AI training and inference using Delta Lake, AWS S3, and metadata cataloging tools, ensuring high-quality, compliant, and scalable data access for model development.

• Enhanced continuous evaluation pipelines with golden datasets, automated testing, and human-in-the-loop validation to monitor accuracy, bias, latency, and safety across all deployed AI services.